Rationale: Although the methods of economic evaluation are widely established, very few evaluations of public health interventions exist and existing reviews suggest that such evaluations raise additional methodological challenges. Objectives: To assess whether the economic evaluation of public health interventions raises additional methodological challenges, to identify those challenges and to suggest ways forward. Methods: Based on existing reviews, 4 methodological challenges were specified; (i) attribution of programme effects, (ii) measurement and valuation of outcomes, (iii) incorporating equity considerations and (iv) identifying intersectoral costs and consequences. The NHS Economic Evaluation Database (NHS EED) was selected to identify economic evaluations of public health interventions undertaken between 2000-2005. These were reviewed, based on the 4 methodological challenges above. Results: The initial search of NHS EED identified 1,264 studies. After excluding duplicate records and studies of screening and interventions delivered in a narrowly-defined clinical setting, 154 studies remained. These offered only a few insights into ways of tackling the 4 methodological challenges. For example, to obtain unbiased estimates of effect, it was possible to undertake RCTs in 38% of the studies reviewed. To extrapolate outcomes beyond the trial or to link intermediate endpoints to final outcomes, similar methods were used as are used in the evaluation of clinical interventions e.g. by using modelling and available epidemiological evidence. The remaining economic evaluations were based on non-randomised studies (31%) and reviews/syntheses (31%). Most of these used some type of model e.g. a decision-analytic model to synthesize data from many sources, or a regression model to adjust for characteristics known to differ between intervention and control groups. In respect of measuring and valuing outcomes, the vast majority of studies did not attempt any valuation, being either CEAs (37%) or CCAs (36%). 27% of the studies reviewed were CUAs, but the valuations concerned were restricted to the health outcomes (states) obtained and expressed in QALYs or DALYs, as opposed to other outcomes beyond health. No CBAs were identified. Equity considerations were rarely mentioned in the empirical studies reviewed. However the search was restricted since NHS EED does not include a search field for equity. The consideration of intersectoral costs and consequences was limited. Although 15% of studies considered productivity costs in addition to healthcare costs, and 9% considered (patients') out-of-pocket expenses, only 4% of studies considered costs in any other sector. Given the relative lack of insights from the current empirical literature, suggestions for ways forward were made following a discussion of the methodological issues, as opposed to current best practice. Conclusions: The existing literature on the economic evaluation of public health interventions leaves a lot to be desired. More innovative use of econometric models is required in situations where controlled studies are not possible. Studies should make more effort to identify intersectoral costs and consequences and to develop ways of comparing the value of outcomes in different sectors, either through new approaches to valuation or developing intersectoral compensation tests. Finally, more research is required on the alternative methods for incorporating equity considerations within economic evaluations
